Code Optimization: Enhanced Configurability for Gradio and LLaVA

[lcolok]

Dear Project Manager,

First and foremost, I wish to express my sincere gratitude for your diligent work on this project. It holds significant meaning to me personally, not only enhancing my skills but also bringing immense inspiration to me.

I have recently made an important improvement to the project, aiming to enhance the configurability of Gradio and LLaVA, making it more suitable for diverse deployment environments, such as Docker containers. I believe these enhancements will make the deployment and management of the project more efficient and convenient in various environments.

The main contents of the improvement include:

Enhanced Configurability: I have extracted several configurable attributes for Gradio and LLaVA, allowing for more parameters to be adjusted during the startup of gradio_web_server.py.

Environment Variable Support: I have added support for the environment variable VIDEO_LLAVA_CACHE_DIR, enabling users to easily specify the cache directory when starting a container. This is particularly crucial for managing model files.

Preservation of Default Configurations: While new configuration options have been added, I ensured that the project still maintains its original functionality when new parameters are not passed, to avoid affecting the experience of existing users.

The purpose of these improvements is to ensure that the project can adapt to a wider range of deployment needs, especially in containerized and cloud deployment scenarios. I believe this is a significant enhancement for the long-term development of the project and the convenience of its users.

I hope these improvements will be well-received. I look forward to your feedback and am willing to further optimize the code based on your suggestions. Once again, thank you for your efforts on this project, and I am delighted to contribute to it.

Best regards